 |
 |
Tracking from lateral motor cortex to the contralateral homologous area is a very hard problem for tractography. The reason is that you have to cross the centrum-semiovale, which contains crossing fibers of the CST and SLF (you need to resolve at least 3 fiber orientations here). If you make it across that, you get funneled into the corpus callosum with a bunch of fibers from all over the brain, and run in parallel with these fibers through the bottleneck of the midsagittal corpus callosum, which is only a few voxels thick at typical 2x2x2mm resolution. Thus, all of the fibers from all over the brain get squashed into a few voxels and it is hard to tell them apart when they spread out on the other end. Then, when you leave the corpus callosum (and this is what kills it even in very high quality data for me) you have to manage to not get swept up towards the superior motor regions which have the most direct route through the diffusion data. Finally, you would have to cross the CST and SLF again, but if you did that okay the first time you ought to be able to do it again. Future work on both diffusion signal modeling and tractography algorithms, along with higher resolution data with higher SNR and many diffusion orientations may be able to solve this problem consistently, however, have only seen hints of this homotopy in the highest quality data with local tractography methods. Peace, Matt. -----Original Message----- From: FSL - FMRIB's Software Library [mailto:[log in to unmask]] On Behalf Of Robert Schulz Sent: Wednesday, January 12, 2011 3:04 PM To: [log in to unmask] Subject: [FSL] problem with interhemispheric, transcallosal tracking in probtrackx dear fslers, currently i am performing intracortical tracking between different motor areas of healthy controls. within one hemisphere, i get quite good results between m1-pmd, pmd-pmv or m1-pmv for instance. transcallosally, there are some connections between sma and pmv/pmd and strongest connections between both sma. if i d like to track from m1 or pmd to the contralateral areas, i haven't got almost any tracks, now after analysing 3 controls. i already checked my exclusion masks, my waypoint mask (CC). I also checked the V1-bedpost orientation maps: could M1 have problems in making the necessary fiber crossings / angles to find its way to the CC area? any ideas? what do you think i could adjust/try? thank you very much for your comments and ongoing help. kind regards, robert